The aim of this paper is to examine the relationship between weather (temperature) and stock market returns using daily data from Portugal; also, to examine whether the temperature is driven by calendar-related anomalies such as the January and trading month effects. Daily financial and weather data from Lisbon Stock Exchange (PSI 20 index) and Lisbon capital for the period 1995-2007 are considered. The paper employs an AR(1)-TGARCH(1,1) model under several distributional assumptions (Normal, Student's-t and GED) for the errors. Empirical results show that temperature affects negatively the PSI20 stock returns in Portugal. Moreover, temperature is dependent of both January and trading month effects. Stock returns were found to be positive in January and higher over the first fortnight of the month. Lower temperature in January leads to higher stock returns due to investors' aggressive risk taking. Further research should investigate the impact of other meteorological variables (humidity, amount of sunshine) and other calendar anomalies on the course and behaviour of major international stock indices using data before and after the recent crisis. The findings are helpful to financial managers, investors and traders dealing with the Portuguese stock market. The contribution of this paper is to provide evidence on the empirical linkages between temperature and stock market returns using GARCH models. To better understand the relationship between the temperature and stock market returns, the paper also examines whether the returns are higher in winter (January effect) and during the first or second fortnight of the month (trading month effect). To the best of the author's knowledge, this is the first empirical investigation on weather and stock market returns relationship for Portugal.
- Statistical distribution
- Stock returns